Achieving Ethical Algorithmic Behaviour in the Internet-of-Things: a Review

IoT devices pose real security and safety risks that go beyond traditional software concerns: The paper highlights that autonomous IoT devices can directly affect the physical environment and people's safety—from hackable vehicles to unsecured consumer devices—making ethical considerations urgent rather than theoretical.

Data collection by connected devices raises significant health and privacy concerns: Health-related IoT systems collect sensitive personal information, and the paper emphasizes that ethical frameworks must address how this data is gathered, stored, and used, particularly given the intimate nature of health monitoring devices.

Algorithmic bias and lack of transparency are embedded problems in IoT systems: The paper identifies algorithmic bias as a key ethical concern in IoT, noting that many systems operate as "black boxes," making it difficult for users and regulators to understand or challenge how autonomous devices make decisions.

Current single-solution approaches are insufficient—a multi-pronged strategy is needed: Rather than relying on one fix (like programming ethics or guidelines alone), the paper suggests that addressing IoT ethical concerns requires combining multiple approaches, including technical solutions, validation methods, developer codes of ethics, and regulatory frameworks.

Accountability gaps exist when multiple connected devices interact autonomously: The paper identifies "uncontrolled cooperation among things" as a concern, meaning that when IoT devices make decisions together without human oversight, it becomes unclear who is responsible when something goes wrong.